Saturday, March 31, 2012

The middle memory

kw: musings, memory, thinking

I love it when a book gets me thinking. I reviewed How to Think Like a Neandertal, by Thomas Wynn and Frederick L. Coolidge, yesterday. A portion of their discussion of Neandertal cognition was memory.

Usually, memory is considered to have two varieties, long term and short term. Short term memory can hold a phone number, or someone's name, or a short list of items, for a few seconds, up to a minute or so. Repetition of a short-term memory, or a sudden shock, can fix it in long term memory, which lasts decades. But another term came up in the discussion of tool making and expertise: working memory. It wasn't really quantified.

Is working memory some combination of short- and long term memory, or something distinct? It seems to work this way. When you are engaged in a task, particularly one which is not wholly familiar, you may have to hold several things in memory for periods of minutes to hours. Yet, once you are done, all the memories fade.

I took advantage of this whenever I wrote a computer program or subroutine. In the very early days, I did what beginning programmers usually do, and built stand-alone applications (today folks just call them apps). They consisted of a stack of FORTRAN code that just did what I wanted it to do, usually. Whether extract certain numbers from a list or calculate a Fourier series, I just set it all up to do what I wanted. While I was writing, particularly in the days of the card punch (no terminal screen to help), I kept the entire program in mind as I spun out the code, line by line.

As time went on, I learned to break up a process into chunks, and write the chunks one at a time. I began to rely on callable routines. Eventually, I could write a program with a general outline:
Program dostuff
Call Askfordata
Call Crankonthedata
Call Printresults
Stop
End
But this just deferred the situation. The core of the work was in routine Crankonthedata, and it could be quite involved, with hundreds of lines of code. Sometimes I could break it up further, but there is a limit to "chunking" a problem. Sooner or later, you have to make it all go.

There was another technique I learned to rely on. As problems I needed to program got harder, it would take several days to write a routine. By such a time, I was pre-writing not the program code, but a flow chart of its operation, and spinning out the code for each block of the flow chart as I came to it. Inevitably, something would not work right, or I'd come to a section and find out I hadn't thought it through sufficiently; there were some dangling threads in the logic that I had to tie up. So at the end of the day, I learned to "grok" the whole routine I was working on, sort of hang the whole thing on a mental blackboard, and go to sleep on it. In the morning, I'd have a batch of stuff ready to write, and it could take half the morning to catch up.

By the way, I always wrote lots of comments in my code, something not all programmers do. Do you know why? I found out that if I looked at the routine just a couple of weeks later, I had no idea what was going on, unless I had good comments to help me decipher what I had written! All trace of the memory was gone. Somehow, a detailed memory, with thousands of parts, that I could hold onto for a few days, would still not make it into long term memory.

Perhaps that's a blessing. I wonder whether my brain would be full now, if I could remember every line of code I wrote over a forty year career. As I've written elsewhere, I used to produce 500-1000 lines a week of FORTRAN. With an average usable "year" of forty weeks (nobody works a 52-week year), that comes to a million lines of code or more.

I once met a pianist who claimed to be nothing special in the memory department, but he typically prepares for a concert by learning 100-150 pages of music. He performs strictly from memory. When he is getting ready for the next concert, any pieces he is not repeating have to be learned. The rest he mostly forgets, though he has a core repertory of a few hundred pages of music that he performs more often.

So what is working memory? It seems to be a forgettable memory store with a medium period, up to a few days or a couple of weeks. If things like FORTRAN code were stored as compactly in the brain as they are on a hard disk, my million lines of FORTRAN might total no more than fifty megabytes. These days, that's about a third of a square millimeter on the surface of a hard disk (the disk in my current computer stores 100 Gby per square inch, or 155 Mby per square mm). Somehow, we may store much more than this about many things, but not everything. I am awed.

Friday, March 30, 2012

What is it like to be a cave man?

kw: book reviews, nonfiction, anthropology, archaeology, neandertals

A prefatory note: The moniker "Neanderthal Man" was coined in the late 1800s. In 1901 the Germans reformed the spelling of their language, and since then (111 years already), the appropriate term has been Neandertal. No matter how it is spelled, the word has never included a "TH" sound. In old high German, "th" was pronounced "T".

In a group sitting around a camp fire, would a Neandertal stand out? Would he or she seem odd in any way? Probably. Physically, the person might seem a little short and stocky, but not much out of the ordinary. The behavior, though, might eventually give the game away. He or she would probably quietly watch the fire, take no part in banter or storytelling, and simply look befuddled at most of the jokes. Let someone bringing a load of wood trip and drop it, though, and our friend just might erupt in the loudest laughter of the group.

In How to Think Like a Neandertal, Thomas Wynn and Frederick L. Coolidge discuss in detail ten aspects of Neandertal life and what they reveal about the way they must have thought. Did they have language? Most assuredly. The common ancestor species, Homo heidelbergensis, is thought by most researchers to have had language and, most importantly, a "Language Acquisition Device", or syntax organ, in their brain, at a level not matched by any ape, even though it was probably less complex than the language and language learning facility in modern humans.

Let's compare us modernized Cro-Magnons with Neandertals, physically; not in detail but just gross anatomy. Average height and weight for Neandertal males would have been an inch or two (2-5 cm) shorter, but 30-50 lbs (14-22 kg) heavier, with a distinct barrel chest. A few "Mr Universe" types can lift one end of an automobile; the average Neandertal male could do so. Average height and weight for Neandertal females was very close to average female C-M height, but 20-40 lbs (9-18 kg) heavier. Whether a Neandertal female would have similar secondary sexual characteristics is not known (and this is not discussed by the authors); she may have had a flat chest when not lactating, and maybe not. Though the average Neandertal had heavy brow ridges and a somewhat receding chin, so do a goodly number of modern folk. In the last chapter, the authors speculate that C-M's were dark skinned and Neandertals more pale. Perhaps. Chimpanzee skin is usually pale under their pelt.

You'd be bored by their diet. "What, mammoth again? At least cook it a bit longer so I can chew it!" Neandertals did not have throwing spears, just thrusting spears. They seem to have fought mammoths close-up, skin-to-skin, where the mammoth could not make effective use of its natural weapons. They had to be fast; getting that last five feet closer is the most dangerous moment. The authors mention just once that many spears were sharpened at both ends. I suspect they were also rather long, perhaps two meters or more: You run in close, jab the mammoth to get it to rear up, and jab the butt into the ground so the mammoth will fall on the tip as it tries to come down on you (you gotta be fast!).

Neandertals, like their ancestor species, were expert flint knappers. This took a lot of repetition to learn, so they were probably more focused than most of us. The chapter on toolmaking has fascinating material that shows just how precise they had to be to strike of those big flakes from a "core". But once expertise is gained, making stone tools was cheap and easy, so the availability of good stone was the only thing that limited how many tools someone could make. They seem to have found it easier to make a new tool than to sharpen an old one. There are millions and millions of Neandertal and earlier tools held in museums and scientists' labs around the world.

Did they tell jokes? Could they understand a joke? Maybe, but it is unlikely. They were not innovators to the extent that we are. It takes an innovative mind to balance the incongruity of a typical joke or pun and see its humor. They'd more likely have appreciated physical humor (Three Stooges?). But their ability to focus and learn by watching would have made them formidable mechanics. If you were to raise a Neandertal baby in the modern world, he or she would probably excel at a number of skills that are learned by apprenticeship, including medicine! Calculus 101 might be a hurdle, though… Take someone with great focus, poor to middling social skills in groups larger than about five, and tremendous ability to learn from repetition: A Neandertal might make a very good computer programmer. Just keep them supplied with mammoth-meat pizzas!

Did they dream? Every other mammal does, so why not? But what did they dream? Probably similar dreams to ours: falling, being chased, forgetting something, looking for something. Many dreams seem to be rehearsals for coping with misfortune, or reminding us not to forget something.

I once read that when tribesmen in New Guinea meet in the forest, they first discuss whether they have any common relatives or powerful friends. They are trying to establish if they can afford to extend friendship, or must strive to kill one another. About half the time, there is a fight to the death. This is probably true of many societies called "primitive", for lack of a less pejorative word. Not all the world's cultures are as comfortable with meeting strangers as the typical Westerner. Neandertals would have been a bit more taciturn and xenophobic, suspicious of anyone outside a family group that seldom numbered more than fifteen. They had to have some tolerance for strangers, or some way of establishing trust, because it is likely that adolescents, either males or females but not both, left their home group to "marry out" into another group. There is little evidence that they carried a gift, which is more common among some modern societies (dowries and bride price customs).

Contrary to some depictions, the authors find no evidence that Neandertals had complex burial practices. They seem to have deposited a fresh corpse in a low spot, perhaps scooped out a bit deeper, and covered it with dirt and stones. More of an "out of sight out of mind" maneuver. They didn't always do so, and often lived among the bones of their ancestors, treating them like sedimentary stones, as they treated the bones of their prey; something to be kicked aside when you sit down. And there is no evidence they expected an afterlife.

Could a baby Neandertal learn a modern language? Could a baby C-M learn a Neandertal language (there were probably a great many languages)? Possibly with ease, in both cases, but more likely, with some difficulty. The brains of Neandertals and Homo sapiens sapiens were somewhat different shapes. So were their thought patterns. In either case, there would be concepts that could not be grasped. The authors think it likely that a Neandertal baby would do better growing up among us, than the opposite scenario.

Neandertal life may have been brutal, but they were not brutish. These were not knuckle-dragging ape-men. They were men and women, they stood upright, and in a suit of modern dress, one of them would not seem unusual. The authors have delivered a great contribution to our understanding of a cousin species, which may also have added a little to our genetic endowment.

Thursday, March 29, 2012

The largest homophone set

kw: words, wordplay, homophones

Air, are (a land measure), ere, e'er, err, Eyre, and heir are seven words that are pronounced the same in most regional dialects of English. I collect groups of homophones (also frequently called homonyms), and of the roughly 800 sets I have gathered, this is the largest. Nearly all the sets have two members, and only a handful have four or more. Of these seven, the two spellings that most clearly express the pronunciation for most people are air, rhyming with hair, and heir, in which the h is silent.

These words are not equally familiar, so let's define each:
  • air is the stuff we breathe. The word can also refer to a popular tune, though this usage is falling out of use: "She sat at the piano and played a simple air."
  • are as a measure of land is to be distinguished from the verb of the same spelling, but pronounced to sound just like the letter r. An are is 100 square meters, and the more familiar measure hectare is 100 ares, roughly equal to 2.5 acres.
  • ere is a poetic preposition meaning "before". It is found at the center of the palindrome that expresses a possible thought by Napoleon in exile: "Able was I ere I saw Elba."
  • e'er Is another poetic expression, a contraction for ever. You'll find it in some older songs, and in many hymns, which frequently rhyme fore'er with there or where.
  • err is a verb meaning to make a mistake: "To err is human; to forgive, divine."
  • Eyre is the surname of the heroine of the novel Jane Eyre. The name is sometimes pronounced with a more distinct y sound, but this is rather rare, and tends to make it a two-syllable word.
  • heir is someone who will inherit. While most people assume an heir is a person's child, your actual heir is whomever you designate in your will. You do have a will, don't you? Without one, upon your untimely death, the government of the state or nation in which you live will decide who your heirs are.
Most European languages are like English in being rather poor in homophones. By contrast, Chinese has so many that, if the language were written with an alphabet rather than ideographic characters, it would be almost impossible to decipher. Even Japanese, which started out as a wholly unique language, has adopted so many Chinese pronunciations of common words that some homophone groups in Japanese have thirty or more members. Of course, this means that these Asian languages are much richer in possibilities for punning!

Wednesday, March 28, 2012

Over 50? You need this.

kw: medicine, medical tests, colon health

OK, this may be TMI for some people: This is how a healthy colon looks from the inside. It is my colon, what there is left of it, looking along the transverse colon toward an anastomosis (surgically fitted joint). At the far end, where the color of the wall darkens, is the ileum, the small intestine.

I had a colonoscopy today, with great results. For the first time there were no polyps. My first colonoscopy, in late 2000, found not polyps but a large cancer. I was 53, and I really should have had a colonoscopy at age 50! Then the removal of a polyp would have prevented all that I went through in late 2000 and early 2001. The good news is, more than eleven years later, I am not a statistic but a cancer survivor.

Many folks are squeamish about having a colonoscopy done. "I don't want anyone running that hose up my butt!" If you see a colon fiberscope, it does look a lot like a six-foot length of black garden hose. Of course, it is a lot more flexible, so it can go through the twists and turns of a normal colon without doing harm.

If I recall right, I've had eight colonoscopies, including that first one in 2000. I had surgery to remove a third or more of my colon a week after that one, and then for five years I had one every year. Five years without a recurrence of cancer is considered a cure, particularly with colon cancer. Those five bring the total to six by the end of 2005. Then I had one in late 2008. Polyps were removed in all of these procedures, so I was scheduled for my next one in late 2011. Scheduling difficulties postponed that to today, only four months "late". My next one is scheduled for early 2015. This will probably continue for the rest of my life.

Most folks need to have a first coloscopy at 50, but if your family has a history of cancer, it is better to start at 40. A friend of ours lost her life to colon cancer at age 49. In my family, my father's side is cancer free, even into their 90s, but my mother's side is more cancer prone, and my mother had cancer three times before dying of ovarian cancer at 81.

Don't be squeamish about it. This test is the only one that can absolutely rule in or rule out cancer of the colon. There is a test for blood in the feces that finds some cancers, but only about half the time. I passed several such tests in my forties! Even at age 51, when I probably already had cancer!!

The test itself is no problem. I've had them with Demerol, with general anesthesia, and this time with a "twilight" treatment where I actually felt the procedure, but felt no pain. The preparation is worse, though that is improving also. In 2000, the preferred "prep" was either a bottle of Fleets Phospho-Soda plus half a gallon of water, all in an hour, or a gallon of Colyte, which is so vile I'd die before I'll ever use it again. I've had both! But since 2002 the preferred bowel cleanser is Senna, a plant derivative that comes in pill form. It is OTC. The usual dose for constipation is one or two pills. The "prep" dose is fifteen pills about noon, and fifteen more in the late evening, with a couple glasses of water each time. Drink more if you get thirsty, and in my case, I did. All this is the day before the test. It cleans you out good, so stay near the bathroom.

This morning, my wife took me to the surgicenter, and an IV was put in for the "twilight" drugs. After a while of sitting and talking with her, I was taken into the 'scope room, where three nurses got me ready while the doctor looked over the photos of past procedures (there are a lot of them at this point). The actual scoping took about ten minutes. As I said, I could feel it this time, and there was no pain. The twilight drugs kept me in a quiet mood (great stuff, if you want a real don't-care attitude for a while). I was expecting more, but the nurse said, "We're done. Let's go," and they rolled me back to the room where my wife was waiting. She was surprised that I was awake and talked to her immediately. A far cry from the times I'd had general anesthesia!

At home, I napped a couple of hours, and now I can at least drive a keyboard and keep my spelling on track. It amounts to a day off work (I worked early on "prep" day, yesterday, and started when I got home); a valuable investment in my own healthy future.

Tuesday, March 27, 2012

Balancing OLEDs

kw: analysis, oled, cleartype

I am a strong proponent of frugality with light, particularly with color image displays. Because of the inefficiency of converting an electron beam to light using phosphors, even high-efficiency rare-earth phosphors, a color cathode ray tube (CRT) about 25 inches in size consumes 200 watts. A liquid crystal display (LCD) of similar size, which uses polarized light and polarization-rotating materials, needs about 80-100 watts, a substantial savings. But the two polarizers in the screen each block 35-50% of the light, and the color filters each pass about one-third of the light, leading to a maximum efficiency of 13% (1/3 of the square of 62%). Almost 7/8 of the light created is filtered out.

By contrast, light-emitting diode (LED) emitters produce light of a specified color (or narrow range of wavelengths), leading to a display that consumes only about an eighth of the energy of an LCD display. There are production difficulties with producing a large LED panel, however, and another big factor is the scarcity of gallium. The LED's in flashlights, for example, use gallium arsenide or a gallium compound with indium and aluminum. Thus a number of companies are working hard to develop organic emitters that use no rare metals. Synthesis of light-emitting polymers is costly at present, but once appropriate materials are discovered, the costs should drop rapidly.

The problem is device lifetime. Red is the low-energy end of the visible spectrum, and red-emitting organic LED (OLED) materials have already been produced that are expected to last between one and two million hours. Green emitters have about half that expected lifetime. Blue, being the high-energy sort of light, is much harder to produce. The best materials, driven to comparable brightness as the red or green emitters, last only about 20,000 hours, at which point their brightness has dropped by half. That is only 2½ years if the display is left on.

Red light in the 640-680 nm range has an energy just under 2 eV/photon. Green in the 520-560 nm range has less than 2.5 eV/photon. But blue in the 420-460 nm range has an energy approaching 3 eV/photon, and it is quite a bit harder to manage without breaking bonds in the molecules. Thus, it is largely a voltage problem, but heating makes it worse, so if you don't drive them so hard in the first place, they can last a lot longer.

I know some people in the field, and have suggested a modified array similar to that shown on the right: leave more space for the blue emitter, and drive it only to half brightness. The red, and to some extent the green, can be given less space and driven harder.

There are other arrangements possible. This could lead to an OLED display that lasts 100,000 hours or more overall. That is almost twelve years, continuous. There is one significant problem, however. Messing with the arrangement this way makes ClearType calculations quite a bit harder. If more than one such modified arrangement is used, things will get messy fast!

This shows how ClearType looks on some italic type on my LCD screen. Note in particular the uprights of the "h" and "r". Sending color information to certain pixels rather than simply "on", "off" or "gray", simulates turning on or off, or dimming, the subpixels, and greatly increases the readability of the text.

Though this can be solved mathematically, it isn't a clean solution. Further chemical research will, I am sure, lead to the discovery of a blue emitter that lasts long enough to make OLED TV's and computer displays a reality. Imagine a 40 inch TV that needs only ten watts!

Monday, March 26, 2012

Ape languages evolving

kw: book reviews, nonfiction, animals, chimpanzees, languages

Primatologist Andrew R. Halloran makes no secret of his stance; in a note early in The Song of the Ape: Understanding the Languages of Chimpanzees, he states, "There are few things more depressing than watching Koko: A Talking Gorilla. I encourage everyone to watch the film and draw their own conclusions." (p 78, note) Later in the book he describes the occasion of first observing the AOL "chat" between Koko and her fans, and how it completely turned him toward discovering how apes communicate among themselves when their experience is not distorted by laboratory methods and researchers' preconceptions. He saw no evidence during the "chat" session that Koko understood either the questions or her "answers".

I have been variously enamored of Koko, Whashoe and other apes that have been taught to use ASL (American Sign Language) signs. I was quite touched at the testimony of one young deaf woman who said, "I was conversing with a gorilla in my mother tongue!" Yet the transcripts of such conversations always left quite a lot to be desired. It requires a fertile imagination to attribute meaning to many of the ape's hand signs, that seem random in retrospect. Dr. Halloran is of the opinion that such signing is of no more significance than a circus trick, and nearly all the "meaning" residing in the mind of the researcher.

Starting with the "chat" in 1998, he spent about ten years as a zookeeper in a Florida animal park, a chimpanzee caretaker, working his way through graduate school. Rather than spend his time in libraries, as grad students are typically expected to do, he carried on his own "laboratory" work among the chimps, getting to know them, learning their habits and particularly their sounds, and finally recording and classifying the vocalization patterns of two groups of related chimps that had been moved to live far from one another.

The result of his work is unequivocal. The two groups, in just a few years, began to develop dialects that differed from one another. After the two groups lost contact (they began as a single large group that was split up when a younger male challenged the alpha chimp), some calls stayed the same but others changed. In particular, just a few years later, the repertory of calls being taught to infant and juvenile chimps in each group constituted a language that belonged to that group. The author expects that in a few more generations, the two groups will become mutually unintelligible. These are the hallmarks of genuine language, with a grammar all its own, but unlike any human language.

During most of the past seven million years, proto-humans (and later, humans) and proto-chimps (and later, chimps), developed their communication systems in isolation from one another. On rare occasions that hominids and apes met, they hunted each other for food. Communication was the last thing on anyone's mind. In central Africa today, chimpanzees are still hunted for "bush meat". It is only in a few Westernized societies that people have developed an interest in communicating with them. Dr. Halloran outlines the work of Yerkes and others who tried teaching chimps to speak, or to use "keyboard" languages such as Yerkish, and finally seemed to have a measure of success with ASL.

Many claim that ape utterances are not language; that they lack grammar or some other feature of human language and are thus not "up to the level" of genuine language. I see only a little merit in such a contention. Near the end of the book, there is a list of a dozen or more utterances that seem unambiguous among a particular group of chimps, with meanings such as "Stop it!", "I am happy with you", or "I found food!". I get the impressions that all the utterances that have been so deciphered are phrases, not words as we understand them. For example, there is a sound meaning "Come here!" that may be a neologism by one of the chimps. It is the only sound listed that has a single "phoneme", a raspberry blurt. There is no utterance yet deciphered that means, "I am coming," so it isn't known whether the phrases are composed of sounds we might call words. However, there are dozens of sounds that were recorded that have yet to be deciphered, and among them there may be some words.

This points up the fact that cross-species communication is hard. Dr. Halloran has not tried to "talk" to the chimps using their sounds, or at least, he does not report doing so. I suspect if he were to do so, it would be similar to my first visit to Japan. I knew a few words and phrases, and in a train station I asked a porter "Benjo wa doko desuka" (Where is the toilet?). He replied "Kore wa rōka ni arimasu, which I didn't understand, but fortunately he also waved down the hallway (rōka, I learned later), where I found the restroom. When asking a question in another language, it is best not to sound too fluent! Better to sound clumsy, and they'll answer more simply. Who knows whether chimps would be so polite to a human?

So far, the author has produced a phrase book that helps him understand certain utterances by two groups of chimps. It is of no help generally, because each group will have its own dialect. There is no more a single chimp language than there is a single human language. I suspect we will find it is equally so for dolphins (and perhaps certain other whales) and perhaps some birds (I am thinking of Alex and other African gray parrots that seem to understand at least some English grammar. I wonder what parrot grammar is like, and how we would find out. Alex has died; will others of his species act as interpreters? Would they have the patience to answer our questions? A parrot's attention span is shorter than any two-year-old's).

Dr. Halloran is not alone. He cites the work of a number of researchers, primarily "on site" primatologists, that have worked to determine the rules of chimpanzee behavior and how the sounds they make relate to their actions. I for one am convinced: An ape's language may be simpler than a human language, but is a language nonetheless. It is a necessary part of the social structure of ape families and groups. Knowing this is so takes the shine off all attempts to teach them any human language. We will do best to learn how they communicate among themselves, and take the onus upon ourselves to communicate with them in ways they find most natural.

Friday, March 23, 2012

When machines can talk, will we know it?

kw: book reviews, nonfiction, artificial intelligence, contests, turing test

How soon will it be that, when we talk on the telephone with a computer, we won't realize it? At present, the experience ranges from frustrating to unsettling, even when it isn't too hard to find what you called for. (If you hate to talk to computers, the source of this image, dialahuman.com, has lists of numbers that are answered by people, for many companies.)

If a computer-generated persona one day makes it through the Uncanny Valley and can converse without creeping us all out, it will have passed the Turing Test. For those who haven't heard, the Uncanny Valley is a phenomenon first reported by Masahiro Mori: when an object is very similar to human or animal, but not quite, it disturbs us more than something a lot less naturalistic. Thus, Pokemon figures are cute and funny, but the 3D-animated figures in the film Polar Express were just unhuman enough to bother a lot of people. For example, they didn't blink in a natural way, if at all.

How about if your only interaction with something is a character-based terminal? Words only, like text messaging. At what point will you find the interaction creepy, and at what point will you be unable to determine whether the something is a human? Actually, considering that some humans are decidedly creepy, a computer-driven teletype doesn't have to cross the valley all the way.

I've known about these issues for a long time. However, I never heard of the Loebner Prize before reading Brian Christian's book The Most Human Human: What Talking With Computers Teaches Us About What it Means to be Alive. The author participated in the competition in 2009 as one of the "confederates", humans who compete with the computer programs to convince the four judges that they are human. The title of the book is that of one of the awards.

The scoring, which I haven't figured out yet, assigns a "humanness" score to each of the contestants, known only by a number to the judges. The chatbot that achieves the highest score gets the award for "Most Human Computer". The confederate who achieves the highest score gets an award for "Most Human Human". Mr. Christian won the latter award in 2009.

Should any chatbot achieve a score of 0.5 or greater from any judge, thus beating out at least one human, it has gone a long way toward winning the prize. In 2009 none of the chatbots did so, but in prior years, scores approaching the threshold of "more human than the humans" have been achieved. Once a chatbot convinces a majority of the judges that it is more human than at least half the humans, it will win the prize, and the contest will be terminated. (Now, that's a pity. At that point, things just begin to get interesting!)

This would be a much more significant achievement than the defeat of Garry Kasparov by Deep Blue in 1997 or last year's victory at Jeopardy! by Watson. Computers excel at the heavy computation needed to play chess or to look up trivia. In each of these cases, the great majority of the work by their creators went into the real world knowledge needed to decipher and rank the input from the outside world. Faced with a more prosaic requirement, such as imitating the nematode Caenorhabditis elegans as it navigates in the soil, I don't think either Deep Blue or Watson would be up to the task. Not just because they weren't programmed to do so, but because no computer has yet been successfully programmed to model the nematode (See NemaSys at U of Oregon for an effort that seems to have foundered about ten years ago).

The book takes us on quite a journey, as the author prepares himself for the competition. Although he had been encouraged by many to "Just be yourself", he judged it unlikely that a person simply running on autopilot would be much different from a machine. Thus, he prepared, mainly by interviewing people who have been in the AI field, former "confederates", and psychologists, to learn how we are human and the ways we might be mistaken for nonhuman.

One achievement many people gain is some amount of familiarity with the piano. We struggle with scales and exercises and etudes, trying to commit them to "muscle memory", which is really brain memory, but of a different sort than remembering someone's name or face. A very well-trained pianist can play most pieces of music perfectly, at full speed, upon first sight. This is called sight-reading. An "ordinary" church organist has to be able to do so, particularly in churches that take requests from the congregation. Yet this is when the most highly-trained person is most machinelike, for a computer programmed to read music and play a piano can do so instantly, without error. It seems the ten to twenty years of practice have produced a piano machine in the pianist's nervous system. But a computer pianist cannot do what the human does easily: adjust tempo and volume to the congregation's singing.

This opens a window into the present lack in AI: senses that are as powerful as our own. The human visual system, for example, has about 140 million sensors in each eye (4% are color sensors) and 500 million neurons in the visual cortex to interpret what the eyes send in through the optic nerves. The best computer vision system so far devised has a few million pixels in each detector and a processor that is the equivalent of half a million neurons or so: roughly equal to the vision of a damsel fly. But a real damsel fly outperforms the computer as it navigates in three dimensions. We hear sounds that are interpreted by some 100 million neurons in the auditory cortex. A similarly-sized system interprets touch and other sensory receptors in our skin that number in the millions.

Just sitting in a chair, as your bottom tires its of contours, provides a sensory distraction that no chatbot feels. A whisper of moving air from a ventilation system may cause a confederate to relax subtly, and may affect his or her interaction with the judge. The chatbot can rely only on what its programmer put "in there" as it reacts to the judge's queries.

During months of preparation for the Loebner Prize event, the author learned of the history of the soul, as it migrated from the heart (whether considered literally or figuratively), to the abdomen, to the head most recently, and now to nowhere, because today's psychologists deny there is a soul. Are we a unitary being of some kind, with psychological "organs" attached, or are we a composite of those organs and nothing else? Are our memories just intellectual entities "somewhere in there"? It seems unlikely, particularly because when we dream, we sometimes smell or hear or feel things both tactilely and emotionally. Most "bad dreams" evoke fear, but there are also happy dreams, angry dreams, erotic dreams, and weepy dreams.

The author devotes one large chapter to "getting out of book", meaning breaking from convention ("Hi", "Hi. How are you?" "Fine, and you?" and so forth). In a five-minute typed "conversation" there is scant time for niceties. Better to jump right in with, "Hi, I'm Brian. What's your favorite hobby?". That'll get the judge into familiar territory and just maybe strike off a wide-ranging conversation that no computer could match. A chatbot that has been extensively programmed for political discourse might hit it off well with the judge who is a political junkie, but fall apart totally when asked, "How 'bout them Astros?"

Just in the past year, a supercomputer was constructed that has processing power and memory capacity both equivalent to a human brain. It won't be long before such a system will fit in a small room rather than a warehouse. How long will it be until the system can feel the wind on its cabinets, hear and see the whippoorwill, smell a rose, tell the taste of chablis from chardonnay (or from a banana!), and laugh at a groaner of a pun? Will a machine system ever be programmed to appreciate NOW NO SWIMS ON MON or "Madam, I'm Adam"? to design a mobile, given a box of small pieces of driftwood and a spool of wire? or at least, to discuss mobiles and Alexander Calder's other art with some felicity? This makes me hanker, just a bit, to try my hand at being a judge…

Mr. Christian concludes that humans are capable of creating "high surprisal" without seeming wholly wonky. Among my friends, this is called "chasing rabbits": one subject leads to another, and the conversation may range over a dozen subjects in a few minutes. This level of flexibility in a chatbot usually comes off seeming schizophrenic, particularly in one like Parry, which simulates the word salad produced by some paranoid schizophrenics. In others it just seems evasive. There is a technique to changing the subject, and most humans do so rather well. As always, too much of one thing or too much of another will give a machine away. Most of us are better balanced than that.

I found it highly entertaining as the author described some of his internal condition during the competition: his worries, his swings from panic to smugness and back. Could the judge detect this in him? Perhaps. And just perhaps, it was the sway of emotion oozing through the teletype-like interface that resulted in his being judged the "Most Human Human" for 2009.

Thursday, March 22, 2012

Extracting interest

kw: complexity theory, images, games, analysis

The Shannon Number, Claude Shannon's 1950 estimate of a lower bound on the number of possible games of chess, is 10120. More recent estimates indicate the actual number is at least 10123, or 1,000 times as large. This is interesting, in that there are "only" about 5x1046 possible board positions. Each position has many possible next moves, so the branching at any point is quite dense.

Very, very few of these games are "good" games, meaning they show competent play, though this can be hard to discern. And whether competent or not, very very few are "interesting". For example, there are huge numbers of games that end abruptly with any of several "fools mate" positions, and a comprehensive computer search would no doubt uncover a great number of "total idiot's mate" positions. Then there are those that quickly lead to a stalemate. The number of possible "interesting" games is impossible to determine, but even if only one game in a quadrillion is interesting, that leaves 10108 interesting games. Even the square root of this, 1054, is a number greater than the number of games that will ever be played on Earth, prior to the loss of its oceans to solar heating about a billion years from now.

Around each possibly interesting game, there will be a number, perhaps a large number, of similar games, that differ by a move here or there, or a few moves taken in a different order, but that in the main cover the same ground and lead to the same final position. There will be a larger number of allied games that lead to a similar but not identical final position.

As I thought about this, I realized that such clusters of similar games are analogous to the cluster of similar images one can produce from any given image by manipulating a pixel here or there, or making wholesale, but subtle, changes. However, the number of possible images, even with a small number of pixels, is much greater than the number of chess games.

Consider this image, a 4x expansion of a familiar icon from Microsoft Windows. The format for these icons is 40x40 pixels, and depending on your screen mode, they can have one byte per pixel (256 colors) or three bytes per pixel (16,777,216 colors). Just one byte per pixel produces 2561,600 possible images, with each possible color palette! That is 103,853. The full color version has 1011,559 possible images. For this image, as for any other, changing the color of any particular pixel by a small amount has no visual effect. Although a computer vision system could immediately tell the difference, the human eye/brain cannot.

Even wholesale changes to the image, such as this shift to a lower gamma, are difficult to discern unless the two images are right next to one another. Let's consider a situation in which we watermark an image by adding one to certain pixel values, leaving the rest unchanged. There are 21,600, or 1.16x1077, possible single bit-level changes in a 40x40 pixel image. If we allow adding 1, 2, or 3 to the low order bits of an arbitrary number of pixels, the number of possible changes for this two-bit-level watermark is the square of the first number, or 1.34x10154, which is 13 quadrillion quadrillion times Shannon's Number. That is the number of effectively undetectable changes to an icon image that could be made. Actually, as other kinds of changes can be made, without visually detectable effect, the cluster of similar images about any given image is much, much larger.

We can try to extract what it is that makes the image
"interesting" by reducing the original image to a 2-level (1-bit) representation. It would take some tinkering to change this image so it looks more like the original, having complete outlines around the pages, for example. So even here, there is a cluster of variations on the theme.

The number of such 1-bit images is the same as the number of single-bit-level watermarks, about 1077, and we can easily see that not all of those will be equally interesting. No matter. There is enough available complexity out there for us to keep devising new icons for the next billion years, without having examined more than a tiny fraction of the possible image space.

There is a lot of world out there. Given the level of complexity available just in a chess game on a board of 64 squares, or in a picture 40 pixels on a side, it is a wonder that anything ever repeats!

Wednesday, March 21, 2012

Another look at a forward-looking book

kw: updated book review

In an earlier post a few days ago, I reviewed Physics of the Future by Michio Kaku. Looking back, I did not do it justice. I was in a bit of a snit over something unrelated, and let it influence the review much too much.

It is a very well thought-out book, in the tradition of earlier books that made predictions that were reasonable and surprisingly accurate. This is because they took advantage of the vision of numerous researchers who were making these futures come to pass. I expect some of Dr. Kaku's expectations to also come to pass. They are well founded in trends we see today.

For example, it is likely that "room temperature" (or higher) superconductivity will be developed within this century. What would this lead to? Present superconductors work at temperatures as high as 133K (-140°C or -220°F). There are claims of superconductivity a hundred degrees C higher, but these materials require high pressure. If superconductivity that remains in effect under a summer sun can be developed, it will probably be found prior to 2100 AD. If, then, such materials can be produced cheaply, some fascinating possibilities arise.

A shallow bowl made of superconducting material will levitate any magnet, up to some limit of magnetic intensity. The intrusion of a magnetic field causes electric current to flow, which excludes the field. This countervailing force lifts the magnet above the superconductor. If a superconductor is created in a large magnetic field, when the field is removed, a current is set up in the superconductor that replicates the field, making it a strong permanent magnet. This latter effect can be used to make super-efficient electric motors and generators.

The former effect can be exploited to levitate things like automobiles, but this assumes that the superconducting material is as cheap as asphalt. Given that, you could pave the interstate highway system with the stuff, put big superconducting permanent magnets in vehicles, and transport becomes very low-cost. Dr. Kaku says it is "almost free" to move something from L.A. to New York along such a roadway. That depends on speed. Air resistance is proportional to the square of velocity. When you have a tailwind, just put up a sail, and you'll soon be zipping along "as fast as the wind". Of course, a quartering wind would require some means of preventing sideslip; the roadway would need to be cupped, for example. But at least half the time there would be no wind, or a headwind, and you'd need a small jet engine to propel the vehicle.

Then there is nanotechnology. What we might call "minitechnology" has led to the 3D printer, that makes shapes from a special plastic, that can be used as a lost-wax (or lost-polymer) process to form a metal part with a complex shape. More recently, 3D printers have been used to produce such things as a running, wind-up clock, though I don't think it runs very long on a plastic spring. The next step is multi-material printing. Then, as the size of the printing "dot" decreases, we could approach total control of material properties by laying down layers atom-by-atom.

I had a "hold your horses" moment when I read about that. It is one thing to move copper atoms about on a carbon surface using an AFM tip, to do things like spell out IBM. Consider that placing each atom requires about a minute's time. Let's let a version of Moore's Law run on this one, and posit doubling the placement speed every two years for sixty years: 230 = 1.07 billion, so the placement time becomes about sixty nanoseconds per atom. Let's see, a billion Iron atoms (atomic mass 56) weighs 93 femtograms, or about 1/10 picogram. That is a cube 227 nanometers across, laid down in about one minute. If you can make something useful, some kind of nanobot, that is about that size, you really have something. But it isn't a useful way to make an automobile or even a compound microscope. As the author notes, getting the atoms to stay where you put them is the real problem.

Is genuine immortality a possibility? More practically, is it possible to eliminate all causes of death other than boneheadedness or accident? Further, is it possible to eliminate aging, so living a very long life doesn't mean being thrust into centuries of misery? Our descendants will need to tackle some very hard problems for this: repair of mitochondrial breakdown; DNA repair; telomere repair that doesn't trigger cancer. Here's what I want: A health span that equals my life span. If the body is subject to all these breakdowns, I want technologies that stave off the painful symptoms of that breakdown until "it is time to go", then a convenient and painless means of closing out life. I am not afraid of death, just of the process of dying.

The most intriguing predictions relate to mental control of our gadgets. With computer methods of interpreting brain waves, we can learn to think thoughts that direct the actions of any machine with an appropriate interface. Some of the prospects are fascinating: Waking up and directing the house to get my clothes out, cook my breakfast, and set my lunch by the door; or carrying on my work by thought control (though I type at 60 wpm). But I am not sure I want someone to come around the corner packing a thought-triggered 9mm Glock, who happens not to like my looks. There are some technologies we need to refrain from developing until psychology is sufficiently developed to identify and repair those who are a bit too Cave Mannish, and have poor impulse control. I don't just mean teenagers, either. I work in a building that has nobody under the age of 35, and in the men's restroom every day I see evidence of "adults" whose mother didn't train them right. I don't know what the women's restroom is like…

Dr. Kaku writes of the developing planetary civilization. This doesn't mean necessarily a one-world government. But some kind of peaceable federation is required. He also introduces the Kardashev typology of world economies, based on total energy use:
  • Type I represents a planetary civilization that uses energy at the rate sunshine falls on the planet. In Earth's case, that would be about 1017 watts, or about 10,000 times current energy consumption.
  • Type II uses about ten billion (1010) times as much energy, or roughly the total energy output of the sun. In Earth's case, the sun produces 2.2 billion times as much energy as what falls on the planet's surface (or upper atmosphere).
  • A Type III civilization uses energy at a galactic level, another factor of about ten billion. In the case of a large galaxy like the Milky Way, there are 100 billion stars or more, but most of them are smaller than the Sun, so the entire galaxy's energy production might be 20 billion times that of the Sun.
On this scale, we can linearize it by taking the logarithm, and the factor of 10,000 mentioned above means that we are a Type 0.6 civilization. It's a long way to Type 1.0 (Type I).

I could go on; it is a big book, full of fascinating ideas. In a few places, the author predicts this or that word will fall out of use. One is cash. If we can make the transition to a truly abundance-based economy, so that anything (within some reasonable limit) can be had for the asking, money loses its power and its value (except for desires outside the "reasonable limit"). Whether this will lead to a world of slackers is uncertain. I'll leave the philosophizing to others for the nonce. I'm just glad I decided to go back and take another whack at the review.

Monday, March 19, 2012

It's in the woods

kw: vacations, sightseeing, photographs

Been in Oregon a couple days, on family business, and the bunch of us went to Camp 18, a restaurant and outdoor logging museum.

Wood carvings abound here. This image is one of their own. Camp 18 is located at milepost 18 on Highway 26, which goes northwest out of Portland to the Oregon coast. It is about an hour's drive out of Portland.

My brothers and I chose the occasion to coincide with our father's ninetieth birthday, so it was a trip for fun as well. We had Saturday free, and had lunch here. We also stopped by Sauvie Island, which someone had said was interesting. That was a bit of a dud, unless you get a kick out of ducks in a swamp.

By contrast, we spent a couple hours wandering around the equipment that is scattered around the land surrounding the restaurant at Camp 18.

What is a logging camp without a portable sawmill? This mill is belt-driven. The camp has several old tractors with take-off pulleys for hooking up a leather belt. I've seen a mill like this in operation. The belt is 20-30 feet long. Click on the image to see a larger version, where you can see that the saw blade's teeth have relief grooves cut; these reduce fouling of the blade. That's one of my brothers moving in for a closeup photo.

Inside the restaurant, one of the first things you see is this quartet of dancing bears, cut with a chainsaw out of logs. We had a great meal there, sitting at a table cut from the stump of a tree that had been nine feet in diameter. I overdid it; their "ordinary" burgers are 1/3 pound, but they had one that was double size. I had to try it. It was eight inches in diameter, on a bun so thick I had to cut the dome off so I could eat it. A couple of my brothers ate the trimmed-off dome.

Outside the restaurant's front door is this welder's whimsy. It was the only bit of artwork that wasn't chainsawn from a log.

This is another iconic bit of equipment: a tree faller (for some reason, loggers don't call it a "feller", which would be more etymologically correct; but what does a logger know of etymology?!) The business end at the left has three clamps to grab onto a trunk as thick as thirty inches. At the bottom, there is a chainsaw with a blade an inch thick, going around a 32-inch bar. Once the faller's clamps grab a tree, the blade is shoved out to saw off the tree. Then the tractor, which outweighs a typical hundred-foot tree by 5:1, simply lays the tree down and unclamps from it.

We also saw several kinds of donkey engine, winches that pulled themselves into the forest on skids; a band saw for cutting logs up to three feet, that had a 2-sided blade (16 inches across) so it could cut logs pushed through in either direction; a portable bunkhouse and camp galley on railroad cars; and a number of rigging poles for lifting and swinging big logs about. We are all mechanical machinery freaks, so it was an occasion for much exclaiming and explaining and arm-waving. A fun day.

Saturday, March 17, 2012

A mere century's prediction

kw: book reviews, nonfiction, science, physics, predictions

If the future were a thing, I suppose we could do something about it. It is, instead, a process based on contingency, a set of outcomes mostly based on current trends, but subject to black swans. By 2001, according to Arthur Clarke's story "The Sentinel" (1951; adapted into the film 2001, A Space Odyssey in 1968), we would have an established presence on the Moon, a "Space Hilton" in a rotating ring-shaped space station, and the capability to send Hal and his pals in a huge ship to Jupiter (or Saturn in the story).

In 2001, none of this came to pass. We had a generation of college graduates who were born after we stopped going to the moon. One small space station had been allowed to crash to Earth, and another one that is moderately useful but is primarily an embarrassment, and an inwardly-focused Western society that seems ever more likely to be swallowed up by a burgeoning Islamicism that nobody dares to challenge openly. In 2100 AD, will the American constitution have been replaced by Sharia law? If so, science will be at a standstill, and world population will be reducing rapidly as we revert to an economic and political order that arose seven centuries ago among a people not noted for their toleration of new things and new ideas.

If all that doesn't come to pass, and more cosmopolitan values prevail, there is a chance that the dreams of Michio Kaku can come to fruition. Author of Physics of the Future: How Science Will Shape Human Destiny and our Daily Lives by the Year 2100, Dr. Kaku is a professor of theoretical physics. He's had the good fortune to get to know many of the scientists who are creating technologies that, they hope, will craft our future: biotechnology, artificial intelligence, nanotechnology and telecommunications.

I enjoyed reading the book, but came away with the melancholy feeling that it reaches too far, and possibly in the wrong direction. The author does anchor himself in the Cave Man principle: inside, we haven't changed much in 100,000 years. Thus, the first century of the Industrial Revolution may have radically improved life in the West, but it also issued in large-scale warfare, starting with the American Civil War, climaxing in the two World Wars, and winding down in the Cold War. The enemies of Western expansionism have adapted by raising guerrilla warfare to a level that nullifies much of the advantage of our high technology. Of course, society has gotten more soft-hearted (and soft-headed), so that we think losing 4,000 soldiers in two wars in the Middle East is too costly. Just as we thought losing 50,000 in Viet Nam was too costly. We lost 450,000 in Europe from 1942-1945, folks, don't forget. Was it too costly? What was the alternative? And we don't turn a hair at the 38,000 (down from 50,000 a decade earlier) that we lose yearly on American highways. We largely ignore the 100,000 (at least) lost to medical mistakes in American hospitals every single year. But a single soldier's death in Afghanistan is considered a tragedy beyond compare.

OK, what did Dr. Kaku write? He predicts that our grandchildren will have godlike powers. They well may. Will they have the wisdom to use them well? The Internet with its advanced searching strategies is the greatest library ever amassed. The majority of searches are aimed at finding pornography. The Cave Man has his dark secrets. Perhaps the Cave Man ought to be eliminated, but take care what you wish for. It is likely that the Cave Man's unique talents brought us where we are, and cannot be dispensed with if we are to continue to progress. We may not like aggressive behavior, but we may not be able to do without it.

Will we become wiser if the expected life span increases by another factor of two or three (into the range of 200 years or so)? If it is accompanied by an extended youthfulness, will women delay having children until they are in their forties or eighties? Will they be willing to keep having periods another fifty years or so? Biotechnology's promoters are promising all kinds of new cures and bodily enhancements, but they all have one of two aims: longer life or more frequent (and better) sex. Just what the Cave Man wants.

How soon will computers or robots be able to do all those things we've been promised for so long: such as household drudgery, dangerous mining, or underwater salvage? Will they ever be able to produce poetry or sonatas or convincing dialog for the theater? These are two ends of the AI spectrum. There is a little rug-cleaning robot already on the market (for a decade or so now). A few programs like Racter produce interesting, if clumsy, dialog. So far, though, no convincing winner of the Turing Test has arisen. I follow the news about autonomous automobiles that don't need a driver. Their recognition tasks are made easier by the use of GPS, but in the absence of satellite navigation, they don't do as well getting down the road as an average cockroach. And Shakespeare and Bach are as yet secure in their positions. My own take is, human-level AI is at least 1,000 years in our future.

Some medical researchers dream of a day when they can send a small army of "nanobots" into a human body, and have a cancer cured, or a bone fracture healed. Materials scientists hope to turn carbon nanotubes into larger and larger single-molecule structures with the tensile strength of diamond, but manufacturable in sizes ranging up to a 33,000 mile tall space elevator. Others are using "quantum dots" and similarly small particles for all kinds of materials with new properties. Will nanotech allow us to produce a contact lens that contains a computer, and can project an image with high resolution right onto the retina. Can you still call that a computer?

I have been jokingly saying for some years that in the near future, having a cellular phone will be considered a right, and that babies will have a phone implant inserted right behind their ear, shortly after birth. It may not be that much of a joke! Connectivity is a dire necessity for my son's generation.

There are a few things I think are quite unlikely. X-ray vision using specially sensitive goggles and a lot-intensity source of moderately soft x-rays, the kind that usually bounce off one's body rather than pass through. These would be similar to the x-ray backscatter devices now being used at some airports. For all but the most frequent travelers, the extra exposure to x-rays at airports may not be a problem, but if x-ray goggles become a consumer item, I'd expect a huge backlash from two constituencies, the folks who don't want privacy invaded and those who don't want the extra radiation. In another place, it is stated that a solar sail might propel a craft to 0.1% of the speed of light. This is possible. However, then it is stated, "perhaps [it will] reach the nearest star in four hundred years." The nearest star beyond the Sun is 4+ light years away. At 0.1% of the speed of light, reaching it would take more than 4,000 years.

There is a good bit of advice, to get work that a robot cannot perform, work that requires pattern recognition and common sense. So far, these cannot be reliably programmed. I would add creativity. There are no robotic Picassos in our immediate future, and possibly not ever. The book ends with a chapter outlining a possible day in the life, circa 2100 AD. Here is where the author needed a collaborator. He simply isn't a fiction writer, and it shows. The rest of the book is well written by contrast. It gave me a lot to think about, an attribute that I prize greatly.

Friday, March 16, 2012

The other way to affect the future

kw: education

Yesterday I analyzed the fraction of one's heritage that comes from more and more remote ancestors. By the time you get 32 generations into the past, the amount is effectively zero in terms of base pairs of DNA. In terms of genes, fifteen generations suffices to reduce an ancestor's influence to less than one gene. Going forward, it takes siring fifteen children to have some assurance that every one of your genes makes it into the next generation. Now many of us have that option.

There is a better and more beneficial way to influence the future. Teach. Our son is planning to be a teacher. He loved his student teaching semester. I taught college classes in computer science for seven years, and quite much enjoyed that. I usually had four sections with thirty students each, every semester. Just counting the ones that got a C or better, meaning they learned something, in seven years I had an effect, positive I hope!, on more than a thousand young people. But it only takes one…

One student I had my second year of teaching the Comp Sci class, visited me a few years later. He'd dropped out of college the year after taking my class, to earn money for his tuition. He got a job in some minor role, but then a programmer at that small company left, and they asked around the office if anyone knew how to program. He volunteered, and within a few months, had been transferred into a computer center position, earning quite a bit more than I was as an adjunct professor (I was still a graduate student). Man, that made my day!

Thursday, March 15, 2012

Not much king left

kw: genealogy, analysis

Caucasian Americans have a fascination with European royalty, and I am no exception. In my own genealogical studies I find it interesting that I am descended through at least three lines, and possibly four, from either Edward I or Henry III. The two of them, father and son, are descended from Charlemagne, sixteen generations back from Edward, who is twenty-two generations back from me via one of the lines.

I began to wonder, just how much of Edward or Charlemagne, or anyone else for that matter, remains in my genetics, this many generations removed? For I have half my father's genes, and half my mother's (plus or minus some statistical variation), one quarter of each grandparent, and so forth. How much of my genome is likely to be descended directly from an ancestor 22 or 38 generations removed?

Ignoring minor variations caused by descent via a male versus a female line, the process of assortment during meiosis, that produces eggs and sperm, is rather blocky and imprecise, and determines what gets included in a particular gamete. There are three figures of interest here:
  • The total human genome consists of 3.2 billion base pairs for a woman, 3.1 billion for a man (the Y chromosome is shorter than the X).
  • The coding portion of the genome is just under 1.5%, or about 48 million base pairs.
  • There are 25,000 or fewer protein-producing genes.
Now I can pick a few ancestors and see how much "influence" each one has on my genetics.
  • John Kirkman and his wife Mary Boyd were immigrants from Ireland. They are four generations removed from me. The factor I'll call Descendancy Fraction (DF) is 2-4, or 1/16, or 0.0625. Genes are usually not chopped during assortment, so 25,000/16, or about 1,560 genes, came to me direct from each of them. Of coding base pairs, I have 3 million from each, and of total base pairs, about 190 million.
  • In the twelfth generation I find Peregrine White, the child born aboard the Mayflower in Plymouth Harbor, and his wife Sarah Bassett. The DF here is 2-12 = 1/4096 = 0.000244… The amounts I have from either of them amounts to just six genes—or enough fragments to make up six—, 11,700 coding base pairs, and about 780,000 total base pairs.
  • Edward I and his wife Eleanor of Castile, 22 generations back: DF = 2-22 = 1/4,194,304 = 2.384x10-7. Here the number of genes is less than one: 25,000/4 million means there is 1 chance in 160 that any of Edward's genes has found its way to me. Coding base pairs? Eleven (plus or minus one or two). Total base pairs: about 760.
  • Now for the real fun: Charlemagne, 38 generations back. DF = 1/2.75x1011 = 3.64x10-12. Here we'll just check total base pairs, and it comes to one chance in 86 that even one base pair from any part of Charlemagne's genome has made its way into mine!
I find that amusing. When my mother was a member of a genealogy club, and our probable connection to Charlemagne was discovered, everyone was so thrilled. If genes have anything to do with who we are, however, there is little chance that such remote ancestors shape any part of what we are. It is still true: you are what you make of yourself.

Now, briefly, let's turn the calculation around. I have one child. A friend of mine is one of fourteen children (he has only one child). Each child bears half the genetic component of each parent. But the assortment process is quite random, so that if you have two children, there is a 75% chance of any particular gene, or any particular bit of your genome, being passed into the next generation. The more children you have, the higher the percentage goes.

For both my friend and myself, half our genes have been sent into our offspring. But for his parents, it is a different story. The DF works in this direction, to calculate the fraction not passed forward. His father's DF, for fourteen children, is 1/16,384 or 0.000061… That means that, at most, one or two of his 25,000 genes was not reproduced into his children. The same for his wife. You can say that everything they are has been sent forward a generation. My friend's portion of his parents' genetics is, however, only 50% passed forward, but of course nearly all his siblings have one child or more, so a large proportion of the legacy of their parents lives on. Having children is still the most reliable way to have an impact on the future.

Wednesday, March 14, 2012

Do I shorten the branch?

kw: genealogy

You'll need to click on this image to see the 1200-pixel original, to make any sense of this. It is a clip from the Ancestry.com display of my family tree.


This shows four generations of ancestors of one Lydia Lamb, a great-great grandmother of mine, followed by four generations of ancestors of Dorcas Gayer, a great-great grandmother of hers—that is eight generations back from myself—and back to Dorcas Gayer's great-great grandfather Thomas Gayer, another four generations, which is twelve generations removed from me, and in ten more generations we reach Edward I Plantagenet, King of England.

The central figure in this diagram is William Gayer, Dorcas's father. A good many descendants of these Nantucketers claim Humphrey Gayer and Jane Sparks as William's parents. Their ancestry is well attested. It is known that they had a son William, but there is some controversy over whether this is the right William. There are two other William Gayers mentioned in documents of the middle to late 1600s, and some contend that one of them is the real father of Dorcas, while others say that the William Gayer, son of Humphrey, is further unknown and the antecendents of all other William Gayers are unknown. Makes life interesting.

The current evidence is, I'll admit, quite circumstantial. It is likely to remain so. This is one of four links in my family tree to English royalty of the 13th and 14th Centuries, and the only one that I can't pin down with certainty. But I do like to retain a link to Ed the First if I can. His engineers produced the first large-scale trebuchet, a weight-driven siege engine. Named War Wolf, it reduced several castles in Wales in wars to solidify the English claims therein. That makes Edward a bit of a bully, but I am quite fond of trebuchets. I've built a few smaller ones, including one that throws a golf practice ball no more than twenty feet. I have used it in a few speeches about the history of warfare.

As for William Gayer? I have a note in my tree that this branch may stop with him. For the nonce, I'll leave it at that. There is always a chance that letters between him and his parents will be found, clearing up the mystery.

Tuesday, March 13, 2012

Bricks, turned

kw: reminiscences, reading

When I was quite little and learning to sound out words, one day I came across a brick sidewalk. A few of the bricks had a word on them. One looked like this:

I sounded it out: "Simons". Great!

I knew a kid whose family was named Simons. Were they in the brick business?

Before long, I came across the next brick with a word on it. It looked different:

I sounded it out: "Snowis". I knew "snow" but this was new.

I wondered if the brick company was named "Simons and Snowis", and decided to look up "Snowis" in the phone book at home.

I turned around to go home, and saw the second brick, and was astonished to see that now it read "Simons"! Feeling a bit odd, I went to the other brick, the one I saw first. Now it said "Snowis".

Finally I realized that they all said just "Simons". It was just which way you came at them.

Some years later I learned about symmetry, and how some words and other objects look the same when rotated. In 1961 I figured out that it was a year with a rotational number, and that the next one would be 6009. That was also the year that I saw a sign that can be rotated half a turn, and reads the same:
NOW NO
SWIMS
ON MON
(Seen at a swimming pool near my home.) These are related to palindromes (Such as "Able was I ere I saw Elba"), but I don't know a term for them.

Monday, March 12, 2012

The hard and the really hard

kw: computers, software, programming, artificial intelligence

I noted earlier the report that a computer system now exists which exceeds the processing power and memory capacity of a human brain. It just needs about nine million watts of electricity to run. However, if things proceed into the future as they have in the past, in thirty years such capacity will be available in larger desktop personal computer systems, and in a further thirty years, in a pocket device, a successor to the smart phone.

There are good reasons to think that future progress may not follow the trend of the past half century or so. Moore's law may be running out of steam. There are several versions of the "law", actually a well-defined trend. The original trend identified by Gordon Moore in 1970 states that the number of devices on a CPU chip tends to double about every two years. In 1971 the 4004 CPU had 2,300 transistors on-chip. In 2011 a 10-core SPARC processor had about 2.6 billion. That is a factor of 1.13 million in 40 years, or just over 20 doublings. So that element of the law has been working just fine. I wonder, though, whether just another ten doublings (a factor of about 1,024) can be accommodated: 2.7 trillion transistors on one chip? On a watch-sized chip (4 sq cm), that is 150 square nanometers per transistor, or a feature size in the 10-12 nm range. That's where it gets hard to keep electrons going where they are supposed to, because of Heisenberg uncertainty.

Other elements? Performance does show signs of hitting a limit. Let's look at a fifteen-year span that is well studied. In 1994 the first Intel Pentium chip was introduced. At 75 MHz, its benchmark speed was 12.2 MFlops. Seven years later, the Pentium 4 ran at 1.7 GHz and benched 152 MFlops, 12.5x faster. From 2001-2009, CPU clock rate didn't quite double, to 3.07 GHz in turbo burst mode in a Core i7, but the benchmark (per core) increased to 667 MFlops, an increase of 4.34x, mainly due to better architecture. The benchmark doubling time in the first seven years was 1.9 years, while in the latter eight years, it was 3.8 years. In the 2006-2009 time frame, doubling time was more like seven years. But now the norm is four, six or eight cores on a large die, making single-thread codes less relevant. I don't expect single-core benchmark speeds to much exceed 1,000 MFlops for some years to come.

To me, all this means that getting the power of the brain into a watch-sized hunk of silicon or a successor material is going to take longer than we might predict, based on the past fifty years of computer hardware history.

There is a second hurdle in the way of getting useful work out of all that power: software development. Do we want a silicon brain to run the same way our lipid-based brain does? It seems a silly idea to me, but not to many proponents of artificial intelligence. Some people are saying that the Watson supercomputer, by winning two days of Jeopardy!, has passed the Turing test. Not really; nobody was trying to make it fool us into thinking it was human. It won a specific kind of trivia contest, as a machine, using machine methods rather than human ones. It was a successor to the Deep Blue chess match against Gary Kasparov. The computer didn't try to behave as a human would, nor was it in any way disguised. Neither system could navigate its way out of a crowded living room (were it mobile).

I don't have a definite figure, but IBM seems to have spent half a billion dollars developing the software code that makes Watson's hardware a Jeopardy! wizard. It will cost dozens of millions more to re-purpose the Watson hardware into a medical diagnostic machine, because of course, diagnostic medicine is not a trivia game, though it does require the marshaling of numerous loosely related facts.

Watson's software is on a par with an operating system. Even your telephone has an operating system. The popular Android OS for smart phones, according to a recent article, has 12 million lines of program code (plus 5 million lines of comments), in forty programming languages and scripts. Roughly speaking a "language" is converted into machine code before use, while a "script" is interpreted from a more human-readable version each time it is used. The compiler for a language is comparatively small: 150,000 lines of code in the case of the Perl compiler. The real heavyweights are full-scale OS's for computers: Linux has 200 million lines of code and Windows 7 is in the 100 million range (Vista had 50 million).

Here is where the kind of CPU you are using has some influence. Much of the code of an OS is in assembly code, and a "line" of assembly code does more on an Intel CPU than on one designed to run UNIX or Linux. So that 100 versus 200 million difference is smaller than it looks.

What does a line of code cost? It depends on the kind of code, but IBM long ago found that a "good" journeyman programmer could write and debug three lines of code daily. A small number of superprogrammers (I was one for thirty years) can do ten to 100 times as much code writing. In FORTRAN, I typically produced 50-100 lines per day. In assembly code, I produced half as much. During the last years I was an active programmer, I earned around $20 per hour, but my work cost my company $50 per hour with overhead, or $400 per day, so a line of my code cost in the $8 range. The larger teams of programmers needed for huge projects like Windows 7 typically include very few superprogrammers, so even with more efficient methods of code generation that are possible using "Visual" languages, a line of code costs $50-100.

Put the figures together. It cost close to a billion dollars to develop Android, and ten times that much to develop Windows 7. That's why Microsoft has to charge $100 to $400 for a copy of the OS, and hope to sell 100 million of them. The first 70-80 million copies just pay the development costs.

Now, consider the human brain. To duplicate all its functions might take billions to trillions of lines of code, if we go the software development route. 'Taint gonna be cheap! Of course, as with an OS, you only have to do it once. But the romantic notion that a lone programmer somewhere will develop a "soul in silicon" is just not in the cards. One of my colleagues was ten times as productive as I was: 500-1,000 lines of good FORTRAN daily (that's a lot of typing, each and every day). So a million lines of code would take him 1,000-2,000 work days. That's four to eight work years. Ten such programmers could produce Android in ten years or less. The actual Android crew, numbering much more than ten, took two years. I tip my hat to them.

Now that we're on the verge of software projects that might be of human-brain scale, can it be done? First, you have to know what you actually want. What would success look like? Right now, if we wanted to start programming "consciousness", we'd be in a position like these folks:

You lot start coding…
…I'll go find out what they want.



There are ten thousand or more studies of what consciousness is. They can't even agree on two or three basic rules to help them recognize consciousness when it appears. Philosophers have been arguing this for centuries (30-40 of them), without producing anything a computer programming team can use as a target. It is going to be an emergent property of some collection of parallel processes, not parallel as doing the same thing, but each set doing something different. But there is no agreement on what are the necessary processes and which ones are simply tools used by a conscious being.

There is not even agreement about whether a physico-chemical body is required. Our brain's operation is strongly affected by hormone levels, and it may be that "our" kind of consciousness (I am including all mammals and birds here) is intimately related to the body's responses to environment, via its chemical cues. I suspect our real "brain" is not just the 1.4 kg of gray+white matter inside our skulls, but includes the other 40+ kg of the body, or at least the 5-10 kg that comprise our nervous lashup plus our endocrine system. I suppose from the total brain's point of view, most of the body is a support system. But the endocrine system may turn out to be essential for any sort of consciousness that we can understand well enough to converse with.

Oh, there is so much to learn, and lots of eager folks trying hard to learn it. It is fun to watch, even though I am pretty much on the sidelines these days.

Some funny ditties

kw: wordplay, verse

I saw this when I was a grade schooler. To figure it out, you have to try to read it aloud, rather quickly:

Seville dar dago
Tousin busis inaro
Nojo demstrux
Sumit cosun sumit dux

Ponder this a while, if you haven't seen it. Translation provided below. Meanwhile, here is one I saw in college. With the exception of one rebus, it is also phonetic. I have to use an image of the whole thing, because there isn't a way to conveniently insert rebuses into text in this scripting tool.


This one may be easier to figure out than the other, at first. Once you get either one, you'll suddenly see the other. So, in case both stump you. . .



First:

See, Willy! There they go,
Thousand buses in a row.
No, Joe. Them's trucks.
Some wit' cows and some wit' ducks.



Second:

'Em are spiders, 'em are.
'Em are no spiders! Oh, yes, 'ey are!
See the eyes? See the beady, beady eyes?
'Em are spiders, 'em are!

Friday, March 09, 2012

Disease or decision?

kw: book reviews, nonfiction, sociology, shoplifting

One of the strongest and most damaging social trends of the past fifty years has been the tendency to make everyone into some kind of victim. In fact, making the perpetrators of crimes into victims has progressed to the point that the only ones having no rights are the actual victims of the crimes!

Sometimes a book comes along, and when I finish it I am at a total loss; what do I think of it? Have I learned something useful? That's how I feel about The Steal: A Cultural History of Shoplifting by Rachel Shteir. It is a well-written and engaging book, but I find myself wondering what conclusions were reached, or indeed, if any could be reached. The author treats of her subject in four areas: its history, its dimensions (variants), its pathology (or not), and its possible remedies, though the appropriately numbered Chapter 13 is titled "The Disease is Incurable".

I reckon that every one of us has either shoplifted or has a close friend or relative who has done so. It is likely that most of us know someone who steals regularly, though we don't always know it. Firstly, it is one of the most under-reported crimes. When the store guards do make a catch, they seldom do more than recover the merchandise and emptily threaten the perpetrator. Strangely, the stores seem to have a sense of shame about it, and don't want thefts publicized. Shouldn't the shame all go to the shoplifter? But perhaps the store doesn't want it known just how easy they were to steal from.

Secondly, it is pretty easy to get away with. Only a small fraction of shoplifting incidents are detected, even by stores with plenty of guards and secret shoppers and CCTV cameras and "tattle tag" detectors. So far, no technology known can recover a signal through a foil-lined bag, and an experienced thief can fool watchers with sleight-of-hand maneuvers learned by every illusionist. In the shoplifting arms race, the advantage has always been with the lifter.

Now, the modern social victimization trend has come along and, even though kleptomania was first discussed a century ago, it is on the verge of being put into the diagnostic manual (DSM V or perhaps VI). There are 12-step programs for "shoplifting addicts", and various other self-help programs all based on either a disease model ("You were born this way") or a victimization model ("You steal because your father didn't spend enough time with you" or "You get a thrill from stealing, that is supposed to come from being praised by your friends"). I suppose, now that portable MRI machines are being developed, someone will be able to measure the mental state of a willing shoplifter as an incident progresses, and determine definitively if it resembles getting high on heroin, or skydiving, or whatever. I'll make a prediction. No doubt, some kind of thrill is involved, but I believe the perpetrator never loses decision-making capacity. Time will tell.

While researching and writing the book, the author encountered a number of colorful characters. Shoplifters who claim to make a living from it; some who "keep their hand in" because of the pleasure of getting away with it; advocates who claim "the man" is stealing from society anyway, so shoplifting is a way of evening up accounts. She also met store guards (AKA Loss Prevention agents or LP's) with lots of training (rare), some with at least a little, and a number who'd learned strictly on the job. She met retailers who were willing to prosecute (also rare), and all along the spectrum to those who seem to feel more shame than the shoplifter.

Is it really true that, without shoplifting, the price of almost everything would be 10% lower? That is one figure bandied about. Does the average store really suffer 10% (or more?) "shrinkage"? I was once told by a traffic officer that, if a law is ignored by 15% or more of the population, it isn't worth trying to enforce. That is why Highway 95 between Philadelphia and Baltimore, and the New Jersey Turnpike, both have regular traffic flow of about fifteen miles per hour over the speed limit. So many people are speeding that the police don't take notice until they perceive someone going at least twenty mph over.

Shoplifting is similar. It is so prevalent that most stores and most police organizations grudgingly content themselves with "keeping it down to a dull roar", so to speak. They only go after a small number of more flagrant cases, some combination of theft-addicted celebrities and high-volume professionals.

I suspect it will take more resources than we are willing to spend to make much of a dent in shoplifting. Will storekeepers a century hence have a brain scanner in the store, so as to detect "dangerous thoughts" before a person even tries to steal? There'll surely be some method or technology for getting past even that. The thieves among us are a combination of the truly bad, the more prosaic "mostly good" folks who sometimes do bad things, maybe some who are genuine victims of a mental disorder, and some unknown proportion of rule-keepers who can't imagine shoplifting.

Ms Shteir leaves such questions open.

Thursday, March 08, 2012

Fleurs 2012

kw: local events, flowers, springtime

Yellow and orange crocuses began blooming the first week of February around here, almost a month early. The first purple crocuses showed up a couple days ago, and some early flowering trees are putting out their first blooms: One kind of flowering cherry, and quince. My pink dogwood looks like its buds are ready to pop, and the white one will surely bloom right after that. Tulip and daffodil leaves are up, but no flowers yet. Curiously, I don't see much sign of Forsythia blooms yet, though I have seen them bloom in February once or twice.

The first time we visited this area seventeen years ago was also an early Spring. We came in late March, and all the flowering trees were in full bloom, a spectacle usually reserved for the second or third week of April. That first day my wife turned to me and said, "I could live here." Of course, the following January we were under almost three feet of snow, and wondering if we'd been so lucky to move here after all.

This makes me wonder what this Summer will be like, and the Winter to follow. The Farmer's Almanac (4 month predictions free to registered folks) is predicting a rather mild March-July 1 season, with only one 90°+ heat wave in June. Lots milder than the past Summer. I guess I rather like La Niña, compared to El Niño. It also means I'd better save up for the extra heating bills of the opposite season, due in 1½ years, and an extra-hot cooling season to follow. Meanwhile, I'll enjoy the flowering season while I can!

Wednesday, March 07, 2012

Would YOU eat there?

kw: cooking, hygiene

I have heard a number of times, and heard again this morning, someone say on the radio, "If you see what's going on in a restaurant kitchen, you'll never eat there." It is a bit like the statement, "Politics is like sausage: You don't want to watch it being made." It also reminds me of the summer I worked in a meat packing plant. It made it hard for me to eat meat until after I left. But I have a stronger stomach now.

The fact is, if you really watch what goes on in your own kitchen, would you be willing to eat the food? Most of us are careful enough, but there are plenty of things we do that are riskier than the standards require of restaurant kitchen workers. Do you wear latex or nitrile gloves to prepare food? Do you wash things as frequently as they do? Do you have a 5-second rule for things dropped on the floor (the restaurant doesn't: into the trash it goes!)?

We've all been told, "Don't leave food out for more than two hours at room temperature." Every church potluck lunch I've been to, the people bring hot food at 9:30-10:00 AM, let it sit on the counter until Noon or 12:30, (some of it is re-heated on a steam table or in a microwave), and it has all been sitting there for two-and-a-half to three hours before any of it is served. Nobody gets sick, though.

Then there's Thanksgiving. Most Thanksgiving dinners I've been to, or hosted, the meal is served pretty promptly after cooking, but then people hang around and "graze" for four to six hours. Sometimes more.

I don't know about you, but I have to have a bit of confidence that my immune system is up to the task, considering that it evolved during the 100,000 years when there was no refrigeration, and some foods were "left out" for days. I know what goes on in a Chinese kitchen—a great many of my close friends are Chinese—but that doesn't keep me from greatly enjoying the occasional pig-out at the Chinese buffet place, or eating at the church's potluck.

Nerve diversity

kw: experiences, dentistry, medicine

I had a permanent crown fitted on a lower molar yesterday, then spent much of the day in a daze, while the anesthetic wore off. I have a rare condition that makes it very hard to "numb" the molars in my lower jaw (the mandible). The nerve to the tooth's pulp enters the jaw far, far back behind the bend of the bone, where it is nearly impossible to reach with a needle full of xylocaine. In most folks (98-99%) the nerve enters just in front of the wisdom tooth, near the tip of its root.

As a result of this unusual innervation, the dentist first has to numb the gums and the tissue at the back of the mouth, then with a much longer needle, try for the nerve. He also used a faster-acting anesthetic in a ligament injector, which forces the stuff between the tooth and the ligament supporting it, so that some will reach the nerve where it enters the tooth—you hope.

The whole side of my head got numb, all right, but when it came time to grind off the temporary crown's cement, I could still feel it. I didn't tell the dentist, because he'd already spent half an hour getting me anesthetized. And we'd been through this before, when he ground off the tooth's enamel and formed a stub for mounting the crown in the first place. I toughed it out then, and did so this time. I could tell that the tooth was partly numbed, but I could still feel the work going on.

When everything was finished, I told him that if I ever need a root canal on a lower molar, it will have to be done under general anesthesia. I told him he had been able to remove most of the sensation, but not all, so I could bear it. Then I went home and to bed. A two-hour nap later, a little sensation had returned, so I had miso soup and some yogurt, neither of which required chewing. I wasn't confident I could eat without chomping out a piece of cheek until eight hours after the dental work. Not only had I had 2-3 times the normal amount of anesthetic, but I get over anesthesia slowly.

This isn't the first time I've encountered odd nerve hookups. When I began having cancer pain twelve years ago, I felt it in the center-left abdomen, where the stomach is, and thought it was an ulcer. Eventually, we found out it was appendicitis, from the cancer eating at the appendix (after growing to the size of a fist in the cecum). The locus of damage was eight inches from where I was feeling it. I just have a goofy nerve connection. Well, the cecum, appendix and a foot or so of both colon and ileum are now gone, so I hope the rest of my nerves are working properly. But there are no guarantees!